HVAC system control, and more specifically, multi-input-multi-output (multivariable, MIMO) control of HVAC plant has a significant advantage over traditional sets of single-input-single output (SISO) loops. MIMO control enables the setting of energy (cost) optimal values of manipulated variables (MVs). The drawback of MIMO control is that setup of the controller is significantly more difficult than setup of SISO loops. The setup requires setting relations between all manipulated variables and all controlled variables, and (ii) entering a cost objective function in a suitable form. A typical procedure for identification of these inputs in process industries (e.g. refinery) requires step-testing led by control engineer skilled in advanced control which is unacceptable in case of HVAC systems due to high costs and impact on building comfort during the testing.